1. Field
Example embodiments relate to noise reduction of captured images, for example, column noise reduction of captured images.
2. Description of the Related Art
An image sensor typically detects light reflected off of a subject and converts the detected light into an electrical signal. The reflected light may be detected by a pixel array of the image sensor, where the pixel array includes a plurality of rows and columns of pixels. A variety of analog processing circuitry then processes the electrical signal output from the pixel array. The electrical signal is representative of the amount of detected light. The processed electrical signal is then converted from the analog domain (e.g., an analog signal) to the digital domain (e.g., a digital signal).
One of the challenges faced by conventional image sensors is that of reducing column fixed pattern noise (FPN). Generally, the main sources of noise in a sensor are a column amplifier (CA) and a thresh and hold (T&H) capacitor. Alternatively, the T&H capacitor may refer to a sample and hold capacitor. For instance, a Cameroon sensor may have one CA and one T&H capacitor per column. The T&H capacitor may add two types of noise: column FPN and frame dependent noise. The amount of column FPN may depend on a type of chip being used by the image sensor. The amount of frame dependent noise may vary according to numerous factors such as image exposure, sensor temperature, etc.
Column FPN may generally be reduced for each column by adding an offset to each column. However, temporal noise as well as some column FPN may still remain. In addition, the added column offset generally does not correct for the column FPN added by the T&H capacitor.